Super Simple DIY Li-Ion Rechargeable Charger

[HarryN]

It seems like a great setup would be if the charger could somehow detect the cell direction + / - (before) it starts to charge, and just stop the charge from happening electrically. I have seen mechanical interlocks beaten by people that thought they were doing the right thing.

A secondary over voltage protection / overcharging protection would also seem valuable, such as Doug S proposed. Last but not least, monitoring temperature would be handy, esp if we start using this thing on larger cells, like 18650. At that point, you might be able to get a UL or TUV listing.

 

[djpark]

[ QUOTE ]
HarryN said:
A secondary over voltage protection / overcharging protection would also seem valuable, such as Doug S proposed. Last but not least, monitoring temperature would be handy, esp if we start using this thing on larger cells, like 18650. At that point, you might be able to get a UL or TUV listing.

[/ QUOTE ]

I have not tried Doug_S's FET protection yet. I am still trying to understand the theory of the protection.

But this LTC4054 chip (and other chips I tried) is rather sensitive to the output resistance and any extra (unnecessary) resistance can alter the charging current which will cause the chip functioning differently. Having a diode or a fuse in series is already bad idea.

There are other chips monitoring the cell temperature, this chip doesn't. But it does monitor the temperature of the chip itself and if it is too hot, it reduce the charging current.

-- dj

 

[andrewwynn]

I have to post a pic of my charger:

that is a 10mm disc it sits on. Almost hard to believe that's a complete LiON charger (just add battery and power of course). there are more complete pictures here

 

[djpark]

I needed to make a li-ion charger for my brother as I am sending a Q3 and PR123/R123. So instead of making an AA holder, I decided to mod the JSB123 charger for him. My initial mod was to put a li-ion cell protection chip and it did work, but why not modify it to house a LTC4054 based charger as used by this DIY charger.

As I am sharing the process of mod I did, it will bring some technical information on JSB123 (4.2V) charger and also other things such as LED charging status indicator and FET reverse cell protection. I hope it can be useful for others.


JSB123 4.2V Charger

This circuit is very cleverly designed buck switcher and transisotr based linear regulator combination to produce 4.25V DC output from input 12V (16V+ rectified) DC. The power supply sayd max 600mA and the charger says 350mA charging current. It is about 0.5C for R123, but the charger actually delivered 200mA only. Both battery holders are joined parallel, no individual control.

The green led is tied permanently to the output (on all the time) and the red led is turned on by a PNP transistor by sensing current passing through a diode. So at starting of the charging the red led is bright and as the charging current drops, the red led gets dimmer.

This is typical charging process. Since the charger output is just 4.25V, it may not trigger the over-charge protection circuit like Pila, so it takes a lot longer time to charge and you may not know when the charging is completed.

I did a few changes to the existing circuit to accomodate the LTC4054 charger chip. I didn't have time to prepare the detail circuit of the JSB charger change, but if many cpfers want, I may start a seperate thread about the JSB charger.

- changed input polarity protection diode from parallel to serial to the DC input
I wonder why it is designed this way.

- added 10k resistor to the feedback divider, so the DC output is now 4.65V instead of 4.25V
The reference voltage is 1.25V so it was easy to calculate.

- added 2 3.3 ohm (1.65 ohm) to the current sensing resistor hoping the max current will jump
I managed to get 300mA max but no more, maybe the inductor or the main power supply capacity?

- connected 15K resistor from PNP transistor base to LTC4054 charge status output pin
Since the LTC output is open drain, it is just so nice to find a PNP transistor based led driver already built in so that I can source the current to the bi-color led instead of sinking.

- added a capacitor and resistor to the DC output of the linear regulator
There is about 2V drop by the 2 transistor regulator and 1 current sensing diode and I find something funny which I can't understand with my limited knowledge. When the power is on, the buck chip didn't oscillated and the it did only when there was a slight output current draw. Somehow the LTC4054 chip didn't trigger the regulator and I had to add a 10K resistor to the regulator output to draw some current and it is stable.


LTC4054 Charger

Since there is some space on top of the existing board, I fixed all the components on top, the spider style.

I didn't have 3.3K SMT resistor, so hooked 3 10K resistor to get 300mA max constant charging current. Also added Vishay SI2314EDS N-channel FET as reverse cell protection as suggested by Doug_S, it works!

A typical charging of JSB123 with this charger is as follows. No major big deal, but just works as expected. Compared to the above charging data.

Charging Status led

Many cpfers asked me how to determine if the charging is complete with LTC4054 charger since the led seems on all the time.

The charging indicator pin of LTC4054 is open drain and it works like mechanical switch to allow current pass through, it does not provide current by itself.

While it is charging, the pin will sink as much as 20mA and the led is very bright. When the charging is complete, it still let 20uA current pass through instead of total cut off. The total cut off is done to indicate the shutdown status when the output voltage is higher than input such as battery is in but no power supply connected.

Even with small current such as 20uA, a bright led will still show some light and it may be mistaken as led on with charging. One way to overcome this is to use a dimmer discrete led display which will not show light at low current.

I and my buddy cgpeanut (Roberto) have discussed about the desire to have red light for charging and green when charging is completed. We went further using a PIC micro-P to monitor the charging current and the battery voltage to indicate different charging statges with different led light pattern or color. I will write a program for PIC to do this sometime later.

Then I got wild to think of using the PIC to control the whole charge process without using a charger chip. It is possible and also feasible to make real intelligent charger. But I recognize it only suitable as school project, not really for a commercial project. There it goes!

Meantime, here is a PNP transistor based high side switch to source the current to led based on the signal from the LTC4054. It is especially suitable to provide current to common cathode bi-color led.

R1 (330 ohm) is used since it is already in the original charger and it gives very nice brightness red light, it may be increased.

R2 (2.2k) is also already there, sos didn't want to change. But the R3 (15K) is what I put and I find it working very fine to clearly distinguish between charging and complete. You may want to change value between 10K and 20K for best result depending on the transistor and R2 value.


Reverse Cell Protection

This N-channel FET based protection circuit is made by Doug_S and MrAl has drawn the circuit together with explanation of theory of working. Since I am not the one to use this charger, but my brother, so I thought a reverse cell protection is important.

As suggest, I used very low Rdson FET Vishay Si2314EDS and built in to the charger. Boy, it was really hard to buy this FET. Thanks to Jay for helping me to get.

The reverse cell indicator led is built into the holder with 100 ohm current limit resistor. This low value gives quite bright red light even with much depleted cell when inserted reverse. It takes about 20mA from the fully charged R123 when inserted reverse without power connected.

Now the time has come for real test!

I connected power supply and with my fingers crossed, I inserted a fully charged R123 reverse. Both charging indicator and reverse cell indicator led were lit. The current measurement from the battery is 50mA only. Thanks God!

I shorted the both output terminal of the charger and measured the current and got 30mA. So total 50mA is correct.

The LTC4054 has output short circuit protection and I presume 30mA limited current is controled by the chip. Some brave souls may want to try this without the FET protection.


Done

It is done! I like it very much, but it is good to know my brother is going to enjoy it. It charges both 14500 and R123 with 2+ hours to complete.

I will update the main post with the new led indicator and fet protection when I find time.

-- dj

 

[andrewwynn]

BEAUTIFUL work DJ..

I am just about to mod one of those infamous 'exploding chinese chargers'..

It has two bays for R123.. and i'm going to make a dual channel charger .. it's tedius to have to make sure voltage is the same before charging pairs.. or not being able to pop on a second while one is charging by itself. In any event.. a lot of what you did here is similar to what i will do.. I want to get reverse protection in.. and a fancy LED feedback like tri-color red/green/yellow..

-awr

 

[djpark]

1) Most China made chargers I've seen including JSB123 charger have 2 battery slots joined together. So if you put 2 cells with different voltage, one battery will charge the other one before both get charged (I presume).

If you want to charge multiple cells with only 1 charging circuit, you can use LTC4411 chip for each cells. It is a low loss ideal diode and it switches Vin to Vout depending on the voltage of Vout pin with 140 mili-ohm FET.

So this chip can be used to charge multiple cells from a single charging source -- charge whichever cell with lower voltage till every cells have same voltage, then start charging together.

Likewise, the same chip can also be used for load sharing from multiple batteries of difference capacity as well.

Another good point of using this chip -- since it is also an ideal diode with very low drop, it will work as reverse cell protection perfectly. No FET needed, you may want to compare the price of this chip and very low Rdson FET.

$0.02 worth of consultation!

2) LTC4054 goes into trickle charge mode when the battery voltage is below 2.9V. That means it provides 10% of the programmed constant charging current. For the same reason, it allow only 10% of the programmed current when the output to battery is short-circuited.

My above experiments confirms that 30ma (10% of programmed 300mA) plus 20mA to led brings total 50mA of current to the battery when the battery is inserted reverse.

It would be interesting to try to put a battery reverse without FET.

-- dj

 

[andrewwynn]

the charger i have they are wired in parallel.. i will cut them apart.. i was just going to make two chargers in one.. double charger chips.. they are only like $1.88.. i'm about to make another order of 75 or so.

I love the idea of that chip though.. (load share).. that is much more clever.

so.. you are saying to put the 4411 chip in series with each battery... and try and see if that alone will provide reverse protection? I think i like the idea of trying that.

Ok.. just checked.. the LTC4411 is more expensive than the charger chip. I pay about 32 cents for the .045ohm Rsdon FET i'm using in my driver and buy them by the 100s, so i can build a dual channel driver with 2 4054s and two of the FETS for just about $4 in chips.

I want to figure out a cool way to implement a 2 or 3 color LED for charging status..

something like:

RED = reversed
YELLOW = charging
GREEN = charged.

I have dual npn chips that are handy.

I'm used to charged = off so it wouldn't bother me if 'green' was 'ready' and 'off' was charged.. but maybe 'dim' green was power on..

what is the mechanism that kills the chip if bat is reversed?

-awr

 

[djpark]

I tend to go on a simpler way to display the led -- 3 independent leds in this case.

green - power on
blue - charging
red - reverse cell

If you want to be fancy with bi-color led, you can monitor the charge pin voltage and determine the charging current as % of the max constant charging current programmed, and display leds according to the charging current with different color or pattern.

1) red blinking - trickle charge (10% current)
2) red firm - constant charging stage (or above 90% current)
3) orange (red + green) firm - constant voltage charging stage
4) orange blinking - near completion
5) green firm - charge completed
6) green blink - no battery

Alternatively using a single color led (say blue), set the led output to 1 Hz PWM with duty cycle according to the charging current ratio. So during constant current charging stage, it will be firmly lit and during the trickle charge, 0.1 sec on and 0.9 sec off. Likewise for 50% charging current during the constant voltage charging stage, 0.5 sec on and 0.5 sec off.

Another $0.02 consultation.

Have fun!

-- dj

 

[andrewwynn]

hey i really like those two cents. I was going to use blue in the nano chargers. but YOUCH didn't really want to buy $100 worth of 5mm LEDs..

I like the idea of green blink for 'ready to go'.. and green solid is 'go'.. i think i'll be happy with yellow for charging... but i really like the progressive charge..

I'm thinking of this: how about a segmented LED.. where the segments light up inversely proportional to current (so easy to get off the prog pin).. It would 'fill' like a gas gauge... i would just have to figure out a reasonable way to do it..i do have some pics and a programer in the house collecting dust.

-awr

 

[djpark]

I have been busy with POP driver and decided to take a short break to cool down the head by making a few chargers.

The basic design is still based on the same Linear LTC4054 single cell li-ion charger controller, but now it has more indicator and I thought I would make a new post, here.

Simple DIY Li-Ion Rechargeable Charger Part-2

-- dj

 

[JamesWong]

Hi again,

Here's the solution for one cell protection.
We'll have to give the parallel cells more thought
i guess. If they are wired into a pack there's no
problem however.

Here it is:

Take care,
Al

I have modified this circuit. The new circuit will better protect reversed battery insertion.
When the battery is reversed, the LED will light up. And the current drawn from the battery is only the current for the LED, no excess current. So the battery will be safe when inserted for unlimited time.

This circuit has not been tested in real life, just been simulated by SPICE. Should be 99% work.

 

[MrAl]

Hi James,

Oh ok, looks good. I guess you simulated it too then?
Build one up yet?

 

[AilSnail]

I made a charger using the TI bq24200, which also works. Followed the instructions in the datasheet.
I was specifically looking for some chip that would not fault when the input is a weak solar cell array (10x15cm). Haven't tried it with the solar cells yet.
With some other chips you can adjust the timer, and some have power good status output, which would help to let me know whether the battery is charged or it is just too low power to charge, or if there is a faulty connection.

 

[petebarchetta]

Does this particular chager have the ability to cahrge more than one cell in parallel? I have two laptop battery packs giving me 12 cells 18650's.
I'm building a chargeable cell of 12 in paralell and output of 2x6 cells giving a 12v output. i have the output sorted, but cant find much in the way of charging more than one cell at a time.
THe cheapo ebay units hoof out from usb, but charge 1 cell, i cant tell if they have the "smarts" in them to detect over charging or are just based in timed charge.

I also have a LI-ion charger for my bike lights 4x18650 cell in parallell driving two cree, but thats just a wall wart and i dont want to sacrifice that unit as its fully sealed.

any help would be good

 

[xeeshan]

hay, i cant download the designs and cant open the next page where you write click here, any Problem in my account or from your side, please correct me.


Thanks